Infrastructure-free Localization of Aerial Robots with Ultrawideband\n Sensors

Abstract

Robots in a swarm take advantage of a motion capture system or GPS sensors to\nobtain their global position. However, motion capture systems are\nenvironment-dependent and GPS sensors are not reliable in occluded\nenvironments. For a reliable and versatile operation in a swarm, robots must\nsense each other and interact locally. Motivated by this requirement, here we\npropose an on-board localization framework for multi-robot systems. Our\nframework consists of an anchor robot with three ultrawideband (UWB) sensors\nand a tag robot with a single UWB sensor. The anchor robot utilizes the three\nUWB sensors as a localization infrastructure and estimates the tag robot's\nlocation by using its on-board sensing and computational capabilities solely,\nwithout explicit inter-robot communication. We utilize a dual Monte-Carlo\nlocalization approach to capture the agile maneuvers of the tag robot with an\nacceptable precision. We validate the effectiveness of our algorithm with\nsimulations and indoor and outdoor experiments on a two-drone setup. The\nproposed dual MCL algorithm yields highly accurate estimates for various speed\nprofiles of the tag robot and demonstrates a superior performance over the\nstandard particle filter and the extended Kalman Filter.\n